This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gebert, C. A. Articles by Waxman, D. J. Search for Related Content PubMed PubMed Citation Articles by Gebert, C. A. Articles by Waxman, D. J. Pubmed/NCBI databases Gene GEO Profiles HomoloGene Nucleotide Protein UniGene Substance via MeSH Medline Plus Health Information Dietary Proteins

Regulation of Signal Transducer and Activator of Transcription (STAT) 5b Activation by the Temporal Pattern of Growth Hormone Stimulation

Division of Cell and Molecular Biology Department of Biology Boston University Boston, Massachusetts 02215

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Plasma GH profiles, intermittent in adult male and continuous in adult female rats, respectively, activate unique patterns of gene transcription in male and female rat liver. Pulsatile, but not continuous, GH exposure activates liver STAT5 (signal transducer and activator of transcription-5) by tyrosine phosphorylation, leading to nuclear translocation, and is proposed to play a key role in GH pulse-regulated male-specific liver gene expression. The mechanisms underlying the GH pattern dependence of STAT5 activation are presently investigated using a rat hepatocyte-derived cell line. Rat GH stimulated tyrosine phosphorylation followed by serine or threonine phosphorylation, leading to activation of the DNA-binding activity of STAT5b, the major STAT5 form present in these cells. Maximal STAT5b activation required a full 20 min at a receptor-saturating GH concentration of 50 ng/ml, suggesting that hormone binding leading to receptor dimerization is a relatively slow process. Repeat cycles of GH pulsation led to repeat cycles of STAT5b activation followed by deactivation, similar to rat liver in vivo. Full responsiveness to succeeding GH pulses required a minimum GH off-time of 2.5 h, but was independent of new protein synthesis. Continuous GH exposure led to down-regulation of activated STAT5b, consistent with the desensitization of this GH pulse-activated pathway observed in female rat liver. The rapid deactivation of STAT5b after termination of a GH pulse involved phosphotyrosine dephosphorylation as a key first step and could be blocked by pervanadate, a phosphotyrosine phosphatase inhibitor. Unexpectedly, serine/threonine kinase inhibitors also inhibited STAT5b deactivation. These studies establish that STAT5b is responsive to the temporal pattern of GH stimulation and demonstrate a role for both a tyrosine phosphatase and a serine/threonine kinase in resetting this JAK/STAT signaling apparatus so that it may respond to subsequent rounds of GH pulse activation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
GH regulates the transcription of a wide number of genes involved in somatic growth, carbohydrate and lipid metabolism, and liver function. While some effects of GH are mediated indirectly through intermediary factors produced in response to GH stimulation, such as insulin-like growth factor I (1), attention has recently focused on the JAK family of tyrosine kinases and on their signal transducer and activator of transcription (STAT) protein substrates (2, 3, 4). Six individual STATs, some with multiple isoforms, have been described as being activated by a growing number of cytokines, growth factors, and hormones (5, 6). Three of these STATs are activated after GH binding to the plasma membrane-bound GH receptor in a process that involves JAK2 kinase-catalyzed (7) tyrosine phosphorylation (8, 9, 10, 11, 12, 13). STAT proteins serve as direct signal transducers to the nucleus that can activate transcription by binding to defined DNA response elements adjacent to target genes (14).

One unique aspect of GH action that has emerged from rodent model studies is the responsiveness of target tissues to GH’s sexually differentiated plasma profile. In adult male rats, GH is secreted by the pituitary in an intermittent manner, to give pulses of GH in plasma that peak at 200 ng/ml every 3.5–4 h, whereas in females GH is secreted more frequently, resulting in the continuous presence of GH in circulation at 20–30 ng/ml (15). Pulsatile GH is more effective than continuous GH in promoting weight gain associated with long bone growth (16, 17, 18); however, the underlying mechanisms by which the temporal plasma profile of GH regulates this important physiological response are not well understood. Liver metabolic function, in particular cytochrome P450 (CYP)-catalyzed steroid and foreign chemical metabolism, also exhibits sexual dimorphism in response to the sexually dimorphic plasma GH profiles (19, 20). P450 enzymes 2C11 and 2C12 are steroid hydroxylases that are exclusively expressed in male and female rat liver, respectively, and have served as prototypic examples of sexually dimorphic, GH plasma pattern-regulated liver gene products (21). In vivo models have established that CYP2C11 gene transcription can be induced in hypophysectomized (GH-depleted) rats given GH in a pulsatile manner that mimics the intact male GH profile, whereas the same gene is suppressed and CYP2C12 is activated when GH is given in a continuous, female-like pattern (22, 23).

Pulsatile GH, but not continuous plasma GH, has recently been shown to activate STAT5 in rat liver, suggesting that this STAT may serve as a direct transcriptional activator of male-specific, GH pulse-activated genes such as CYP2C11 (12). Indeed, STAT5 contributes to GH regulation of CYP3A10, which is a male-specific, GH-dependent steroid 6ß-hydroxylase expressed in hamster liver (24). GH also activates two other STATs in liver tissue, STAT1 and STAT3 (11, 13, 25) but, in contrast to STAT5, the activation of these other signaling molecules by GH is largely independent of its temporal plasma profile (13). GH-induced tyrosine phosphorylation of each STAT is followed by phosphorylation of the STAT on serine or threonine residues (13). Tyrosine phosphorylation is necessary for STAT5 DNA-binding activity, but the role of serine/threonine phosphorylation is less well understood. While the hypophysectomized rat liver model used in these earlier studies has proven very useful for elucidation of these responses of STATs in an intact liver and in a physiological context, mechanistic questions relating to the GH pattern-dependence of STAT5 activation, deactivation of STAT5 after termination of a GH pulse, and the apparently slow desensitization of the GH receptor/JAK2 kinase/STAT5 pathway in liver cells exposed to GH continuously (12) have been more difficult to address.

Recently, an immortalized rat hepatocyte-derived cell line, CWSV-1, was described as responding to continuous GH treatment by induction of insulin-like growth factor I, steroid 5-reductase, and GH receptor mRNA (26). In the present study, CWSV-1 cells are characterized with respect to their expression of a functional JAK/STAT pathway that is shown to respond to GH rapidly and in a manner similar to intact liver in vivo. We describe a GH pulse-responsive STAT5 protein, STAT5b, that undergoes both tyrosine- and serine/threonine phosphorylation, and we provide evidence for a role of a pervanadate-sensitive tyrosine phosphatase in the rapid deactivation of STAT5b after termination at a GH pulse. STAT5b deactivation is also shown to be facilitated by a serine/threonine phosphorylation reaction, thereby resetting the JAK/STAT signaling apparatus in a manner that enables STAT5b to respond to a subsequent GH pulse-induced activation event.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
GH Activation of STAT5b in CWSV-1 Cells
CWSV-1 is an adult male rat hepatocyte-derived, SV40-transformed cell line that retains some differentiated functions of normal liver, grows in a chemically defined medium (27), and can respond to continuous GH by induction of several liver-expressed mRNAs (26). To ascertain whether GH activates one or more STAT proteins in CWSV-1 cells, cells were treated with rat (r)GH. Cell extracts were subjected to electrophoretic mobility shift assay (EMSA) using a rat ß-casein gene STAT5/MGF response element probe, shown previously to bind PRL-activated mammary gland STAT5 (28) and GH pulse-activated rat liver STAT5 (12). Whole cell extracts prepared from rGH-treated but not untreated cells showed a single specific band bound to the DNA probe (Fig. 1A, lanes 2 and 3). DNA-binding specificity was demonstrated by the full competition observed with 50-fold molar excess of unlabeled DNA probe (Fig. 1C, lane 6). The DNA-protein complex was fully supershifted with anti-STAT5b antibody but not anti-STAT1 or anti-STAT3 antibodies (Fig. 1C, lanes 9–11), demonstrating that this GH-activated STAT5 protein-DNA complex does not contain STAT1 or STAT3 and likely corresponds to a STAT5 homodimer. EMSA and anti-STAT supershift analysis using the c-fos gene SIE (s-cis-inducible element) probe, which can be used to detect GH-activated STAT1 and STAT3 in rat liver (13), revealed low levels of complexes containing STAT1 and STAT3 in GH-treated CWSV-1 cells (data not shown). Thus, all three STATs are activated by GH in CWSV-1 cells, as is observed in intact rat liver in vivo (13).

View larger version (53K): [in this window] [in a new window]   Figure 1. STAT5b Is Specifically Activated by GH Treatment of CWSV-1 Cells CWSV-1 cells were treated for 45 min with rGH, hGH, PRL, and the G120R hGH mutant at the concentrations indicated, and total cell extracts were prepared as described in Materials and Methods. A, Cell extracts were subject to EMSA with the ß-casein probe, which formed a specific protein-DNA complex induced by rGH or hGH. B, Cell extracts were analyzed by Western blotting with a STAT5b-specific primary antibody. STAT5b-immunoreactive band, designated band 2, was induced by rGH and hGH and corresponds to a tyrosine + serine/threonine-diphosphorylated STAT (see text).1 C, ß-Casein gel shift complex seen in panel A was confirmed as specific by self-competition using unlabeled DNA probe at the indicated molar excess (lanes 6–8) and was identified as containing STAT5 by supershifting with anti-STAT antibodies (lanes 9–11). Untreated CWSV-1 cell extracts are included as negative controls (lanes 1–4). The extracts analyzed in panels B and C were from cells treated with hormone concentrations of 50 ng/ml for 45 min.

Analysis of CWSV-1 cell extracts on 5–10% gradient SDS gels, followed by anti-STAT5 Western blotting, revealed that STAT5 band 0 was comprised of two distinct components, both of which were substantially depleted in cells treated with high GH concentrations (500 ng/ml) and were converted to the more slowly migrating band 2. The intensities and relative levels of these two closely spaced constitutive STAT5 bands were not affected by treatment with several phosphoprotein phosphatases, including calf intestinal phosphatase, a general phosphatase (see Fig. 3A, below, and data not shown). The relationship between these two constitutive STAT5 bands is not known; however, both correspond to STAT5b isoforms, insofar as they both react with the STAT5b-specific antibody used in these experiments (see Materials and Methods). Neither band reacted with a STAT5a-specific antibody, which revealed a single band in both untreated and GH-treated CWSV-1 cells that migrated close to the position of STAT5b band 2 (data not shown).

View larger version (43K): [in this window] [in a new window]   Figure 3. Influence of Phosphatase Treatment on Electrophoretic Mobility of STAT5b: Evidence for Both Tyrosine Phosphorylation and Serine/Threonine Phosphorylation Induced by GH A, Extracts from untreated cells (lanes 1–3) and 45 min GH-treated CWSV-1 cells (lanes 4–6) were incubated with the phosphotyrosine phosphatase PTP1B (lanes 2 and 5) or the phosphoserine/threonine phosphatase PP2A (lanes 3 and 6), as detailed in Materials and Methods, and then subjected to anti-STAT5b Western blotting. PTP1B did not alter the STAT5b-banding pattern in untreated cells (lane 2 vs. lane 1), but it eliminated band 2 and increased the relative intensity (as a fraction of total STAT5b immunoreactivity) of band 1a in GH-treated cells (lane 5), confirming band 2 as containing phosphorylated tyrosine. PP2A treatment decreased band 1a in untreated cells (lane 3), indicating that this STAT5b band is serine/threonine phosphorylated. PP2A also decreased the relative abundance of STAT5b bands 1a and 2 in GH-treated cells with the formation of band 1 (lane 6 compared with lane 4). This confirms that STAT5b band 2 also contains phosphoserine/threonine residues. B, CWSV-1 cells were treated with 500 ng/ml GH for 0, 10, or 45 min (lanes 1–3) or were pretreated with H7 (lanes 4–6) or genistein (lanes 7–9) followed by GH treatment. Extracts were analyzed by anti-STAT5b Western blotting of an SDS gel run under conditions that maximize the resolution of STAT5b bands 1 and 1a (e.g. lanes 2 and 6).

View larger version (47K): [in this window] [in a new window]   Figure 2. Activation of STAT5b Is Mediated by Tyrosine Phosphorylation CWSV-1 cells were preincubated with the indicated kinase inhibitors as detailed in Table 1. GH (50 ng/ml) or vehicle was then added and cell extracts were prepared 45 min later. A, STAT5 activity was analyzed by EMSA using a ß-casein probe, revealing that the GH-induced formation of activated STAT5 complex (lane 2) was substantially blocked by inhibitors of tyrosine kinases (lanes 6, 8, and 12) and by a broad-range kinase inhibitor (lane 10), but not by a serine/threonine kinase inhibitor (lane 4). B, Corresponding Western blot shows STAT5b band 2 is present in GH-stimulated cells that were otherwise untreated (lane 2), but is essentially absent in lanes 6, 8, and 12. The weakened STAT5b band 2 and the appearance of a STAT5b band that migrates between bands 0 and 1a after GH stimulation of H7-pretreated cells (lane 4) is consistent with H7 inhibition of STAT5b band 1 to band 2 conversion (see text). C, Cells were treated with 50 ng/ml GH for times ranging from 5 min to 1 h, then analyzed by anti-STAT5b Western blotting (lanes 1–6) or by immunoprecipitation with anti-phosphotyrosine antibody PY20 followed by anti-STAT5b Western blotting (lanes 7–12). The predominant GH-induced tyrosine-phosphorylated band (lanes 10–12) corresponds in mobility to STAT5b band 2 (lanes 4–6). A weaker band(s) corresponding to STAT5b band(s) 1/1a was present in lane 10, and to a lesser extent in lane 11, as seen on a longer exposure.

Further support for these STAT5b band identifications is provided by an analysis of the STAT5b banding patterns on a high resolution gel, which clearly separates band 1 and band 1a (Fig. 3B). GH treatment initially yielded the tyrosine-phosphorylated band 1, which was subsequently converted to the tyrosine + serine/threonine-phosphorylated band 2 (lanes 1–3). In cells pretreated with H7, the basal level of band 1a was lowered (lane 4), consistent with our identification of this protein as a serine- or threonine-phosphorylated STAT5b form. GH stimulation of the H7-pretreated cells resulted in the accumulation of STAT5b band 1 at 10 min (lane 5), followed by its slow conversion to STAT5b band 2 (lane 6). That the conversion of band 1 to band 2 is slowed down, but is not fully inhibited, by H7 suggests two possibilities: 1) the serine/threonine kinase active on STAT5b band 1 is only partially inhibited by H7; 2) two distinct kinases can carry out the serine/threonine phosphorylation reaction, but only one of these kinases is inhibited by H7 under the conditions of these experiments. The serine/threonine-phosphorylated band 1a also appeared with longer time GH treatment (lane 6). This may arise by phosphotyrosine dephosphorylation of band 2 (c.f., phosphotyrosine dephosphorylation preceding phosphoserine/threonine dephosphorylation, below). Finally, both STAT5b phosphorylation reactions were blocked by the tyrosine kinase inhibitor genistein (lanes 7–9), supporting our conclusion that GH-induced STAT5b tyrosine phosphorylation precedes serine/threonine phosphorylation.

Indistinguishable STAT5b-DNA EMSA mobilities were exhibited by untreated and H7-treated CWSV-1 cell extracts (Fig. 2A). This contrasts with the appearance of a new, lower mobility complex (EMSA complex II) when activated STAT5 present in liver nuclear extracts was converted to its monophosphorylated (phosphotyrosine) form by in vitro treatment with PP2A (13). In agreement with this observation, STAT5b-DNA complex II was not formed after PP2A treatment of GH pulse-activated CWSV-1 extracts but was formed when GH pulse-activated liver nuclear extracts were treated and analyzed in parallel (data not shown). The absence of complex II in the case of CWSV-1 cells may reflect a subtle difference between the STAT5b molecules present in the two systems, or alternatively, may result from the presence in liver but not in CWSV-1 cells of a novel protein that is a component of STAT5 DNA complex II.

Kinetics of GH Activation
The time course of STAT5b activation in CWSV-1 cells was investigated as shown in Fig. 4A. At 50 ng/ml GH, STAT5b activation was faint or absent at 5 min, became detectable at 10 min, and was maximal by 45–60 min. Similarly, Western blot analysis demonstrated the appearance of the diphosphorylated STAT5b band 2 initially at 20 min and then maximally at 45 min (Fig. 2C). The time required for STAT5b activation was progressively shortened as the concentration of GH was increased from 50 ng/ml to 2000 ng/ml (Fig. 4A). Since the dissociation constant (K_d) for the GH-GH receptor complex measured in intact cells is 100 pM (2 ng/ml) (29), it is expected that even the lowest concentration of GH used in these experiments, 50 ng/ml, should be saturating with respect to GH receptor binding. The apparent requirement of a high GH concentration for rapid STAT5b activation could, in part, be due to a low-level, constitutive phosphotyrosine phosphatase that dephosphorylates one of the components required for this response (e.g. JAK2 kinase, GH receptor, or STAT5b itself), thereby antagonizing the effects of GH. This possibility was tested by pretreating CWSV-1 cells with the phosphotyrosine phosphatase inhibitor sodium pervanadate (60 µM for 1 h), followed by GH stimulation. At both 20 ng/ml GH and 50 ng/ml GH, the kinetics of STAT5b activation were not affected by the presence of pervanadate (Fig. 4B). Thus, the comparatively slow rate of STAT5b activation seen at the lower GH concentrations is not due to reversal of the effects of GH by the action of a pervanadate-sensitive phosphotyrosine phosphatase.

View larger version (52K): [in this window] [in a new window]   Figure 4. Kinetics of STAT5b Activation Are GH-Concentration Dependent A, CWSV-1 cells were incubated with GH at the concentrations and for the times indicated. Cell extracts were analyzed for STAT5 activity by EMSA using a ß-casein DNA probe. The STAT5b-DNA complex appeared at progressively earlier incubation times as the GH concentration increased. B, CWSV-1 cells were incubated in the presence or absence of 60 µM pervanadate for 1 h prior to GH addition at 20 or 50 ng/ml. This treatment did not alter the activation profiles seen in EMSA gels with the ß-casein probe.

STAT5b Is Rapidly Deactivated upon Removal of GH, but Can Be Reactivated by Repeated GH Pulsation
In rat liver in vivo, STAT5 is exclusively localized in the cytosol until it becomes activated by a GH pulse, which induces nuclear translocation of the STAT in close association with GH-induced tyrosine phosphorylation (12). In order for STAT5 to respond to sequential GH pulses, as has been shown to occur in rat liver in vivo, a mechanism must exist for rapid termination of a STAT5 response after each GH pulse. This could involve STAT dephosphorylation, perhaps associated with recycling of the STAT protein back to the cytosol, or perhaps deactivation by degradation followed by new STAT5 protein synthesis. We first investigated whether CWSV-1 cells respond to repeated GH pulses by repeated cycles of STAT5b activation. Cells were incubated with GH in a manner that mimicked the physiological, male plasma GH pattern, where GH pulses of 1 h in duration are followed by a GH-free period of 2.5 h (overall 3.5-h pulse frequency, measured peak to peak). The GH concentration was set at 50 ng/ml for these cellular studies to approximate the average GH concentration across the 1-h hormone pulse in vivo (c.f. peak GH concentration in plasma 200 ng/ml) (17). Upon addition of GH, STAT5b activation was faint at 5 min, maximal at 45 min, and still robust at 60 min (Fig. 5A, lanes 2–5). Removal of GH at 60 min (corresponding to termination of a GH pulse) led to a complete loss of STAT5b DNA-binding activity within 30–60 min (Fig. 5A, lane 6; also see Fig. 6A, lane 3), and this was associated with a corresponding disappearance of STAT5b band 2 detected by Western blotting (Fig. 5B, lane 6). Addition of a fresh aliquot of GH 2.5 h after the first GH pulse resulted in the reactivation of STAT5b² (Fig. 5, lanes 8–11). A third pulse of GH resulted in a third cycle of STAT5b activation (lanes 14–17). Thus, STAT5b undergoes repeated cycles of activation and deactivation in response to a male pattern of pulsatile GH stimulation.

View larger version (47K): [in this window] [in a new window]   Figure 5. STAT5b Can Be Activated Repeatedly by GH Given Intermittently, to Mimic the Male GH Pattern CWSV-1 cells were incubated with GH (50 ng/ml) for 1 h, then washed and placed in fresh RPCD medium. After 2.5 h, GH was again added to the culture medium for 1 h. This cycle, diagrammed at the bottom of the figure, was repeated a total of three times. A 60-mm dish of cells was taken to prepare total cell extracts at each of the time points indicated. A, EMSA of the extract using ß-casein probe shows STAT5b activated by each of the three GH pulses. B, Corresponding anti-STAT5b Western blot, indicating the appearance of STAT5b band 2 in response to each GH pulse. The ratio of activated STAT5b band 2 to band 1 is greater after GH pulse 1 (lanes 4 and 5) than after pulse 2 or pulse 3 (lanes 10, 11, and lane 16). Image shown in lanes 1–5 is from the same Western blot but was prepared using a somewhat lighter x-ray film exposure than lanes 6–16.

View larger version (39K): [in this window] [in a new window]   Figure 6. Repeat Activation of STAT5b Requires an Interpulse Interval for Full Responsiveness, but Does Not Require New Protein Synthesis A, CWSV-1 cells were treated with 50 ng/ml GH for 1 h, then washed to remove GH and fed fresh GH-free RPCD medium. After intervals of 0.5, 1, 2, and 3 h, a fresh aliquot of GH was added, as indicated in the diagram. Cell extracts were taken before GH treatment (lane 1), 45 min after the first addition of GH (P1, lane 2), at the end of each of the indicated interpulse intervals (lanes 3, 5, 7, and 9), and 45 min after the onset of pulse 2 (P2) (lanes 4, 6, 8, and 10). The ß-casein probe EMSA shown demonstrates reactivation of STAT5b, which is progressively more robust with longer interpulse intervals. Full responsiveness required a GH-off time interval of 3 h. B, Shown is an EMSA of the effects of cycloheximide (10 µg/ml), administered 30 min after the onset of an initial GH pulse at 500 ng/ml. This treatment did not alter STAT5b responsiveness to a second GH pulse (lanes 3a and 3b, vs. 2a and 2b), indicating that new protein synthesis is not the basis for the interpulse interval requirement shown in panel A. Data shown in lanes marked ’a’ and ’b’ correspond to duplicate tissue culture plates.

Role of Phosphotyrosine Phosphatase in the Deactivation of STAT5b After Termination of a GH Pulse
Conceivably, the requirement of 2.5–3 h for full responsiveness of STAT5b to a second GH pulse (Fig. 6A) may reflect degradation of one or more components of the GH receptor/JAK2/STAT5b pathway, followed by a need for new protein synthesis. To investigate this possibility, CWSV-1 cells were incubated with the protein synthesis inhibitor cycloheximide (10 µg/ml), after which cells were given two pulses of GH, each 1 h in duration and separated by a 3-h interpulse interval. This experiment was carried out at a concentration of GH, 500 ng/ml, which resulted in a near complete conversion of STAT5b to its tyrosine-phosphorylated, activated form (band 2; cf. Fig. 3B, lane 3). Cycloheximide given 1 h before GH did not decrease the intensity of the response of STAT5b to a GH pulse (data not shown). Cycloheximide pretreatment at either 10 µg/ml (Fig. 6B) or 50 µg/ml (data not shown) also did not decrease the response of STAT5b to a second GH pulse. Thus, the comparatively slow resetting of the GH-activated STAT5b pathway does not result from a time requirement for new protein synthesis.

To ascertain whether phosphotyrosine phosphatase activity is required to reset the STAT5b GH signaling pathway, cells were treated with the membrane-permeable phosphotyrosine phosphatase inhibitor pervanadate (33). While pervanadate had no discernible effect on the kinetics of STAT5b activation (Fig. 4B), it markedly slowed the deactivation of STAT5b after termination of a GH pulse. This was evident from the prolongation by pervanadate of STAT5b DNA-binding activity during the interpulse interval after removal of GH (Fig. 7A) and from the persistence of STAT5b band 2 on Western blots of these same cell extracts (Fig. 7B). Incubation of CWSV-1 cells with pervanadate alone induced a very slight activation of STAT5b (Fig. 8, lane 1). This suggests that the prolongation of the STAT5b signal through the interpulse interval in pervanadate-treated cells (Fig. 7A) may be due to inhibition of the deactivation of existing, activated STAT5b molecules, rather than continued activation by a constitutively activated JAK2 (cf. Ref.34). The fact that STAT5b band 2 persists in the presence of pervanadate, and is not converted by a phosphoserine/threonine phosphatase to band 1, demonstrates that STAT5b phosphotyrosine dephosphorylation precedes phosphoserine/threonine dephosphorylation. This conclusion is supported by our finding that the phosphoserine/threonine phosphatase inhibitors, okadaic acid and calyculin, did not block STAT5b deactivation (Fig. 7, A and C) nor did they block the dephosphorylation leading to loss of STAT5b band 2 after GH removal (Fig. 7B, and data not shown). We conclude, therefore, that STAT5b tyrosine dephosphorylation obligatorily precedes phosphoserine/threonine dephosphorylation and is a key step in the STAT5b deactivation pathway.

View larger version (36K): [in this window] [in a new window]   Figure 7. Inhibitors of Phosphotyrosine Phosphatases and Serine/Threonine Kinases Delay Deactivation of STAT5b after Termination of a GH Pulse A, CWSV-1 cells were treated with GH (50 ng/ml) in two 1-h pulses separated by a 2.5-h interpulse interval. Sets of culture dishes were preincubated with okadaic acid, sodium pervanadate or H7 (Table 1), as indicated. Cell extracts were prepared 20, 45, or 60 min after GH addition, or 1 h or 2.5 h after GH removal (interpulse), as indicated. Shown are ß-casein probe EMSAs, which revealed pervanadate and H7 are both effective in prolonging the activated STAT5b signal after GH removal, but okadaic acid is not. B, Western blotting of extracts described in panel A revealed the persistence of band 2 during the interpulse interval in pervanadate-treated cells (lanes 8 and 9), relative to untreated cells (lanes 3 and 4). A weaker band 2 is also seen to persist in H7-treated cells at 1 h (lane 13) but not at 2.5 h after termination of the GH pulse (lane 14). H7 pretreatment was associated with a relative intensification of STAT5b band 1 after GH stimulation (lane 12), consistent with H7 inhibition of band 1 to band 2 conversion. Note that band 2 is present in lane 5 but is too weak to be seen in the reproduction shown in this figure (cf. pres-

View larger version (94K): [in this window] [in a new window]   Figure 8. Down-Regulation of STAT5b Activity by Continuous GH Treatment CWSV-1 cells were treated continuously with 50 ng/ml GH in the presence or absence of pervanadate and H7 (Table 1), as indicated. Extracts were made at the times ranging from 20 min to 24 h and analyzed by EMSA using the ß-casein probe. The activated STAT5b signal was maximal at 45 min and by 2–3 h was down-regulated to 15–20% of it peak value, which was maintained for at least 24 h. Pervanadate prolonged the peak level of activated STAT5b for about 2 h, whereas H7 maintained the peak STAT5b level for at least 4 h.

Role of Serine/Threonine Phosphorylation in STAT5b Deactivation Pathway
GH-induced activation of JAK2 tyrosine kinase is followed by activation of serine/threonine phosphorylation carried out by an unidentified GH-responsive kinase. Targets of this latter kinase activity include tyrosine- phosphorylated STAT5b, as well as STAT1 and STAT3 (13). To investigate the role of serine/threonine phosphorylation in STAT5b deactivation, CWSV-1 cells were treated with the serine/threonine kinase inhibitor H7. H7 significantly inhibited the decline in STAT5b DNA-binding activity after termination of a GH pulse (cf. Fig. 7A, lanes 5, 6; H7 vs. corresponding no drug controls). This protective effect of H7 was less complete than that provided by pervanadate, suggesting that serine/threonine phosphorylation of one or more protein factors (including, perhaps, serine/threonine phosphorylation of STAT5b itself) facilitates, but is not absolutely required for, STAT5b phosphotyrosine dephosphorylation. Similar protective effects, albeit to differing extents, were observed in experiments using structurally diverse serine/threonine kinase inhibitors, including H8 (Fig. 7C and Table 1). The effectiveness of H7 with respect to kinase inhibition was confirmed by its partial inhibition of the accumulation of STAT5b band 2 after GH treatment (Fig. 3B, lanes 4–6 vs. 1–3). Conceivably, a more complete inhibition by H7 of the GH-stimulated accumulation of STAT5b band 2 might have resulted in a more complete inhibitory effect on STAT5b deactivation. Interestingly, in contrast to the inhibitory effect of pervanadate on a second cycle of STAT5b activation (Fig. 7A, lanes 7–9 vs. lane 6; see above), H7 did not interfere with the repeat activation of STAT5b by a second GH pulse (Fig. 7A; note increase in STAT5b activity in H7 samples from lane 6 to lanes 7–9). This is consistent with our earlier observation that although GH induces phosphorylation of STAT5b on both tyrosine and serine/threonine, tyrosine phosphorylation alone is sufficient to activate STAT5b’s DNA-binding activity (13).

Female Pattern of Continuous GH Exposure Leads to the Loss of Activated STAT5b Complex
To investigate whether STAT5b responds in a differential manner to continuous as compared with intermittent GH stimulation, CWSV-1 cells were exposed to 50 ng/ml GH continuously for times up to 24 h. Figure 8 shows that STAT5b was activated maximally during the first 45 min, but within several hours the level of STAT5b DNA-binding activity declined to 15–20% of the peak level. This reduced level of activated STAT5b was maintained for at least 24 h. The same time-dependent suppression of activated STAT5b was obtained at 5 ng/ml GH and at 500 ng/ml (data not shown).

Continuous GH treatment of CWSV-1 cells in the presence of pervanadate delayed the initial phase of STAT5b deactivation by 1–2 h, indicating a role for a phosphotyrosine phosphatase in this deactivation. A role for a serine/threonine kinase was also apparent from the finding that H7 maintained activated STAT5b at close to its initial level for at least 4 h in the presence of continuous GH. The apparent effectiveness of pervanadate for a shorter period, only 2 h after GH addition (i.e. 3 h after pervanadate addition to the cells), reflects the toxicity associated with pervanadate treatment at longer exposure times. Longer term studies on the effects of phosphatase inhibition on the responses to continuous GH could not be carried out due to the toxicity of these compounds, which became clearly evident by 3–4 h in the case of pervanadate and after 5–6 h in the case of H7.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
The present study establishes that a STAT5 DNA-binding protein, identified as STAT5b, can be specifically activated in the rat liver cell line CWSV-1 by GH-induced tyrosine phosphorylation followed by serine or threonine phosphorylation. Activation of this STAT was almost fully reversed within 30 min following cessation of GH treatment, but could be reinduced after a recovery period of 3 h by subsequent pulses of GH modeled on the physiological male pattern of intermittent plasma GH stimulation. Continuous GH exposure, modeled on the female rat plasma GH profile, initially stimulated STAT5b activation, but ultimately led to down-regulation of the STAT5 response. The rapid decline of activated STAT5b after termination of a GH pulse involved phosphotyrosine dephosphorylation as a key first step. STAT5b deactivation could thus be blocked by inhibitors of phosphotyrosine phosphatases but, unexpectedly, was also markedly slowed down by inhibitors of serine/threonine kinases. Both classes of inhibitors also blocked the initial phase of STAT5b deactivation after continuous GH exposure. CWSV-1 cells thus provide a useful cell culture model for investigation of the STAT5b activation pathway and its responsiveness to the pulsatile pattern of GH stimulation that characterizes STAT5b signaling and GH-induced transcriptional responses in adult male rat liver in vivo. This model system also provides insight into the mechanisms that underlie the deactivation of STAT5b after termination of a GH pulse and the down-regulation of the STAT5b response in cells that receive the continuous GH stimulation that is characteristic of adult female rat liver.

GH Receptor Activation in CWSV-1 Cells
GH initiates cellular responses by dimerization of the plasma membrane-bound GH receptor (35, 36), leading to dimerization associated with autoactivation of JAK2 kinase. The G120R mutant of hGH binds hGH receptor with a high affinity (K_d = 0.3 nM), but only interacts with the receptor via site 1 of the GH molecule, thus blocking the receptor dimerization event required for signaling (29). G120R-hGH failed to activate STAT5b in CWSV-1 cells, in accord with its inactivity in other GH-responsive systems (37, 38) and consistent with a requirement for receptor dimerization. It should be noted, however, that when administered to hypophysectomized rats, G120R hGH is poorly bound by rGH receptor and may interact with PRL receptors (39). While GH, but not PRL, is presently shown to activate STAT5b in CWSV-1 cells, GH and PRL both activate STAT5b in the human hematopoietic cell line U17 (40) and in transfected COS-7 cells (41). Non-liver-derived cell lines, especially ones reconstituted by transfection of essential signaling components, may exhibit receptor couplings different from those normally found in hepatocytes. In this regard, the present findings using CWSV-1 cells closely reflect the in vivo rat liver model, where STAT5 (primarily STAT5b) is specifically activated by GH but not PRL (12).

GH binds to its receptor with a K_d of 0.1 nM (2.2 ng/ml) (29), suggesting that the receptor would be saturated in CWSV-1 cells even at the lowest GH concentrations used in this study (20–50 ng/ml). While this was apparently the case (cf. similar maximal level of activated STAT5b at 50–2000 ng/ml GH; Fig. 4A), the kinetics of STAT5b activation were rather slow at the lower GH concentrations (20–45 min required for maximal STAT5b activation at 50–125 ng/ml GH). The GH concentration-dependence of these kinetics was unaffected by the presence of dexamethasone, which in 3T3-F442A cells can down-regulate cell surface expression of GH receptor (32). The kinetics of STAT5b activation were also unaffected by the phosphotyrosine phosphatase inhibitor pervanadate, and thus the slow rate of STAT5b activation at lower GH concentrations does not result from a high basal rate of tyrosine dephosphorylation of one or more components of the pathway (JAK2, GH receptor, STAT5b). These kinetics are similar to those seen in GH receptor-transfected cells (42) and most likely reflect the GH concentration-dependence of receptor binding leading to STAT5b activation per se, which at physiological GH concentrations (50–200 ng/ml) appears to be slow. Indeed, GH activation of liver STAT5 in hypophysectomized rats in vivo is not detected until 10 min after intraperitoneal hormone injection (13).

STAT5b Phosphorylation and Dephosphorylation
In rat liver in vivo, GH-induced STAT5b tyrosine phosphorylation is followed by phosphorylation of STAT5b on serine or threonine (13). GH also induced formation of a tyrosine + serine/threonine-diphosphorylated STAT5b in CWSV-1 cells (Fig. 3; STAT5b band 2), although the presence of a high basal level of serine/threonine-phosphorylated STAT5b in these cells (band 1a) made it difficult to establish whether this diphosphorylated STAT5b resulted from GH-induced tyrosine phosphorylation of preexisting band 1a or whether it reflects tyrosine phosphorylation of STAT5b band 0, followed by a secondary, GH-stimulated serine/threonine phosphorylation reaction, as occurs in liver in vivo (13). Support for the latter hypothesis comes from two observations: 1) STAT5b band 1 appears as a transient intermediate along the pathway that forms STAT5b band 2 (Fig. 3B, lanes 1–3); 2) formation of the diphosphorylated STAT5b band 2 is inhibited by the serine/threonine kinase inhibitor H7, and this results in accumulation of the tyrosine-phosphorylated band 1 (Fig. 3B, lane 5). However, given the high basal level of STAT5b band 1a in CWSV-1 cells, it is still possible that this preexisting serine/threonine- phosphorylated STAT5b form may also serve as a substrate for the GH-induced tyrosine phosphorylation reaction. The source of the endogenous serine/threonine-phosphorylated STAT5b band 1a is unknown; levels of this STAT5b form were greatly decreased in H7-pretreated cells (Fig. 3B, lane 4), but could not be reduced by overnight culture of the cells in the absence of insulin or other additives in the RPMI culture medium (data not shown).

Phosphotyrosine dephosphorylation, rather than STAT5b protein degradation, appears to be the primary mechanism for termination of a STAT5b signal. This is indicated by the ability of the phosphotyrosine phosphatase inhibitor pervanadate to block the rapid deactivation of STAT5b after termination of a GH pulse and is supported by our finding that the recycling of STAT5b and its response to a subsequent GH pulse are unaffected by protein synthesis inhibition. Hepatocytes contain numerous phosphotyrosine phosphatases associated with various intracellular compartments, including membranes, cytosol, and the nucleus (43, 44). The SH2-domain-containing phosphotyrosine phosphatases SHP1 and SHP2, in particular, can play both positive and negative regulatory roles in the action of other cytokines and growth factors that signal via JAK kinases (45, 46, 47) and could participate in STAT5b deactivation by binding to tyrosine-phosphorylated signaling molecules either at the level of the GH receptor/JAK2 kinase complex or at the level of nuclear STAT5b. While the present study shows that STAT5b phosphotyrosine dephosphorylation obligatorily precedes phosphoserine/threonine dephosphorylation, STAT5b tyrosine dephosphorylation appears to be facilitated by a prior serine/threonine phosphorylation step. Whether this involves serine/threonine phosphorylation of STAT5b itself or of some other signaling factor cannot be established with certainty (also see below). Further studies are required to elucidate the mechanisms of STAT5b deactivation, including any regulatory effects that GH may have on the phosphotyrosine phosphatase(s) that contribute to this process.

The kinase that catalyzes the secondary, GH-induced phosphorylation of STAT5b on serine or threonine has not been identified, but could correspond to mitogen-activated protein kinase, which is also activated by GH (25) and can catalyze serine phosphorylation of other STAT proteins (47). Tyrosine phosphorylation, but not serine/threonine phosphorylation, is required for STAT5b’s DNA-binding activity (Fig. 2A). Although not tested directly, GH-induced STAT5b serine/threonine phosphorylation may contribute to STAT5b’s transcriptional activation potential, as has been shown in other STAT systems (48, 49), including interleukin 2-activated STAT5 (50). In addition, the present study suggests that a serine/threonine kinase activity (perhaps the GH-activated serine/threonine kinase that acts on STAT5b) sensitizes STAT5b to deactivation by tyrosine dephosphorylation, as indicated by the inhibitory effects of several serine/threonine kinase inhibitors on STAT5b deactivation. This intriguing observation raises several possibilities, including: 1) the phosphotyrosine phosphatase(s) that deactivate tyrosine-phosphorylated STAT5b may have a preference for STAT5b molecules that are also serine or threonine phosphorylated; 2) inhibition of serine/threonine phosphorylation prevents nuclear translocation of STAT5b, thus shielding the STAT protein from nuclear phosphotyrosine phosphatases that may catalyze its tyrosine dephosphorylation and deactivation; and 3) serine/threonine phosphorylation, perhaps catalyzed by a GH-regulated kinase, may activate the phosphotyrosine phosphatase that deactivates STAT5b. However, the fact that H7 treatment not only prolongs STAT5b band 1, but also prolongs the diphosphorylated band 2 (Fig. 7B, lanes 12 and 13 vs. 2 and 3) argues against the first two possibilities and thus lends support to possibility 3.

A low constitutive activation of STAT5b in the absence of GH could be detected when CWSV-1 cells were incubated with pervanadate, which in other cell models can induce tyrosine phosphorylation of multiple proteins, including JAK2 (33, 34). STAT5b was activated in pervanadate-treated CWSV-1 cells to only a small extent, whereas STAT3 and STAT1 were significantly activated under the same treatment conditions, as demonstrated by SIE probe EMSA with supershift analysis using anti-STAT antibodies (C. A. Gebert and D. J. Waxman, unpublished experiments). This difference between the STATs may reflect the preferential association of STAT3 (and perhaps also STAT1) with JAK2, as compared with STAT5b, which apparently requires a more direct interaction with the COOH-terminal cytoplasmic domain of the GH receptor to undergo activation (51). Whether the low basal level of activated STAT5b in pervanadate-treated cells results from its interaction with a JAK2 kinase/GH receptor complex, or perhaps with another tyrosine kinase, is not known. These hormone-independent stimulatory effects of pervanadate on STAT activation suggest that cellular phosphotyrosine phosphatases may in part serve to reverse the effect of adventitious STAT activation resulting from random association with tyrosine kinases, cytokine receptors, or other signaling molecules, thus maintaining a minimal level of tyrosine-phosphorylated STAT in the absence of ligand.

Response of STAT5b to GH Pulses: STAT5b Recycling
The ability of CWSV-1 cells to respond to repeated pulses of GH with repeated cycles of STAT5b activation and deactivation is consistent with in vivo studies in GH pulse-treated hypophysectomized rats and with the finding that in intact adult male rats nuclear translocation of activated liver STAT5 is intermittent and coincides with the occurrence of a plasma pulse of GH (12). STAT5 is thus activated for about 1 h of each 3.5- to 4-h GH pulse/GH trough cycle. It will be interesting to determine whether the intermittent activation of liver STAT5 leads to intermittent transcription of GH pulse-responsive genes. These genes are likely to include male-specific CYP genes such as CYP2C11, whose transcription is induced by GH pulses in adult male rat liver and is suppressed by continuous GH treatment in adult female rats (22, 23). Indeed, a functional STAT5-binding site has been identified in the 5'-flank of CYP3A10, a GH-regulated, male-specific hamster gene (24). In hypophysectomized rats given pulsatile GH replacement, CYP2C11 expression could only be detected in animals given six or fewer GH pulses per day, corresponding to an interpulse interval of 2.75–3 h. By contrast, treatment with seven GH pulses per day, which results in a shortening of the interpulse interval by only 35 min, fully blocked the activation of CYP2C11 (17). In the present study, CWSV-1 cells also required an interpulse interval of 3 h for complete responsiveness to a second pulse of GH, but in contrast to the CYP2C11 gene transcription response in the in vivo model (17), STAT5b activation could be partially restored with shorter interpulse intervals, with the strength of the resultant STAT5 response being roughly linear in relation to the length of the interpulse interval.

The mechanism underlying this requirement of a relatively long interpulse interval (up to 3 h) for full restoration of STAT5b’s GH pulse responsiveness is unknown. It does not reflect time required for resynthesis of one or more of the signaling proteins involved in the STAT5 pathway, as demonstrated using the protein synthesis inhibitor cycloheximide. Rather, the extended time interval for restoration of full GH responsiveness may be dictated by the time needed for one or more of the following events: 1) recycling of GH receptor back to the plasma membrane after its internalization to various intracellular compartments, such as the golgi and the nucleus (52); 2) phosphotyrosine dephosphorylation of signaling proteins such as JAK2, GH receptor, or STAT5b itself. Such a dephosphorylation requirement is supported by our finding that whereas pervanadate prolongs STAT5b activation after termination of a GH pulse, subsequent GH pulses do not further increase the level of activated STAT5b (Fig. 7A); and 3) translocation of deactivated STAT5b from the nucleus back to the cytosol. That STAT5b is likely to recycle to the cytoplasm after GH pulse activation is indicated by the absence of a protein synthesis requirement for repeat GH pulse activation of STAT5b, even when the cells are exposed to a concentration of GH (500 ng/ml) (Fig. 6B) that results in a near quantitative phosphorylation of the cell’s STAT5b pool (Fig. 3B). On the other hand, STAT5b recycling per se seems unlikely to be rate limiting with respect to a second GH pulse response since the 3-h interpulse interval requirement is also seen in cells treated with GH under conditions (50 ng/ml) where only a portion of the cellular STAT5b pool becomes phosphorylated at any given point in time. Interestingly, despite significant differences in the rate of STAT5b activation at low vs. high GH concentrations, similar maximal levels of activated STAT5b were detected by EMSA under both conditions of cell treatment (Fig. 4A). This maximal level detected at 45 min after hormone treatment was not altered by pervanadate treatment, a finding that is consistent with our observation that STAT5b dephosphorylation is not initiated until 40–60 min after addition of GH to the cells (C. A. Gebert and D. J. Waxman, unpublished experiments).

Finally, continuous exposure of CWSV-1 cells to GH, in a manner that mimics the female plasma GH pattern, led to a down-regulation of the STAT5 response that became apparent within 2 h of continuous GH treatment and was maintained for at least 24 h. This response was substantial, albeit only 80–85% complete under the conditions of these cell culture experiments, and generally correlates with the down-regulation of liver STAT5 in vivo in hypophysectomized rats given GH continuously (12). The persistence of activated STAT5b at a low level for 24 h or longer apparently results from continued GH-induced signaling, rather than from an inhibition of phosphotyrosine phosphatase activity, since removal of GH after either 1 h, 2 h, 4 h, or 24 h of continuous hormone exposure leads to a rapid and complete loss of activated STAT5b (C. A. Gebert and D. J. Waxman, unpublished experiments). Whether the prolonged, low level activation of STAT5b in continuous GH-treated cells is mediated by JAK2 kinase or by another tyrosine kinase is unknown.

In conclusion, CWSV-1 cells are shown to provide a useful cell culture model for study of the effects of plasma GH patterns on the activation and deactivation of STAT5b. Further investigation should help elucidate additional mechanistic details that underlie this pathway, including the role of serine/threonine kinase(s) and the contributions of phosphotyrosine phosphatases to STAT5b deactivation and its recycling back to the cytosol. GH-activated STAT5b analyzed on Western blots can be seen to contain as many as five bands in H7-pretreated cells, suggesting additional complexities in this system that remain to be identified.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Cell Culture
CWSV-1 cells (27) were passaged (1 to 6 split) in 3% FCS/RPCD medium (26), aliquoted evenly to 60-mm tissue culture dishes at a density of 15,000 cells/cm², and incubated at 37 C overnight. The medium was then aspirated and replaced with serum-free RPCD medium. Medium was subsequently replaced every 2 days until the cells were about 80% confluent, when the cells were either passaged or treated with GH. RPCD medium composition was essentially as described (26), with a trace element component as follows: 0.25 nM NiCl_2, 0.25 nM SnCl₂, 0.5 nM MnCl_2, 0.5 nM (NH₄)₆Mo₇O₂₄, 2.5 nM Na₃VO₄, 15 nM Na₂SeO_3, 50 nM CdCl_2, 250 nM FeCl₃, 250 nM Na₂SiO_3. The RPCD medium was made up by combining three parts, comprising 1) the trace element mixture, 2) basal RPMI 1640 medium plus 0.28% bicarbonate and 0.36% HEPES buffer, and 3) BSA (0.08%) containing the following growth factors: linolenic acid, 2 µg/ml; 2-aminoethanol, 4.1 µg/ml; glucagon, 0.04 µg/ml; dexamethasone, 0.4 µg/ml; insulin, 0.06 µg/ml; transferrin, 100 µg/ml, as recommended by Dr. H. Isom, Pennsylvania State University. CWSV-1 cells were stocked in 50% FCS/40% RPCD medium/10% dimethylsulfoxide and stored in liquid nitrogen until used. To ensure a high survival rate, stocks were prepared using cells that were 2 days past their previous passage. Once thawed, CWSV-1 cells retained their morphology and GH responsiveness for about eight to ten passages so long as 1) the cells were passaged before complete confluency, and 2) the cells were in contact with trypsin solution [0.31 g trypsin (Sigma T-8253), 0.125 g glucose, 0.35 g EDTA, 48 mg NaHCO₃ in 1 liter PBS] for no more than 2 min during the splitting procedure.

For GH and/or inhibitor treatments, 1 ml of fresh medium was used per 60-mm tissue culture dish on the day of the experiments. GH (dissolved in PBS containing 0.1% BSA and kept on ice until use) was added in volumes of 10 µl using schedules described in each individual experiment. Data shown are generally representative of at least three independent experiments. Rat GH, hGH, and PRL were hormonally pure preparations obtained from the National Hormone and Pituitary Program, NIDDK. The hGH mutant Gly 120 ->Arg (hGH-Gl20R) was provided by Drs. J. Wells and G. Fuh (Genentech, South San Francisco, CA). GH used in each experiment was rGH unless indicated otherwise. Kinase and phosphatase inhibitors used in this study were used at concentrations and under conditions shown in Table 1. Inhibitors were prepared as 100x stock solutions and stored at -20 C, with the exception of pervanadate, which was prepared fresh daily as a 100x stock containing 6 mM Na₃VO₄ and 5.7 mM H₂O₂ (final nominal concentration, 60 µM pervanadate). Most of the inhibitors shown in Table 1 exhibited little or no apparent toxicity to the cells in the time frame of the experiments; H7, H8, and HA1077 were significantly toxic to the cells after 5 h incubation, while pervanadate toxicity become apparent by 3–4 h.

Total Cell Extracts
Cells were washed once with ice-cold PBS and then scraped with 100 µl of lysis buffer (20 mM HEPES buffer, pH 7.9, 1% Triton X-100, 20% glycerol, 20 mM NaF, 1 mM each EDTA, EGTA, Na₃VO₄, Na₂P₂O₇, dithiothreitol, 0.5 mM phenylmethanesulfonyl fluoride, and 1 µg/ml each pepstatin, antipain, and leupeptin). The crude extract was aspirated ten times through a 27 gauge needle, adjusted to 150 mM NaCl, and centrifuged at 15,000 x 30 min at 4 C. Supernatants were stored in liquid N₂ until analysis. Protein concentrations were determined using the Bio-Rad Dc detergent protein assay kit.

EMSA
Total cell extracts (10 µg) were made up to a total volume of 5 µl in lysis buffer, then added to 8 µl of gel-shift buffer (5% glycerol, 1.25 mM MgCl₂, 625 µM EDTA, 625 µM dithiothreitol, 12.5 mM Tris, pH 7.5) plus 1 µl containing 2 µg of poly(deoxyinosinic-deoxycytidylic)acid (Boehringer Mannheim, Indianapolis, IN) and incubated 10 min at room temperature. Double-stranded, ³²P-labeled oligonucleotide probe (1 µl, 10 fmol) was then added to give a total volume of 15 µl. Incubation was continued for 20 min at room temperature and then 10 min on ice, followed by addition of 2 µl loading dye (30% glycerol, 0.25% bromophenol blue, 0.25% xylene cyanol). The incubation mixture was loading onto an acrylamide gel (5.5% acrylamide, 0.07% bis-acrylamide (National Diagnostics, Atlanta, GA) in 0.5x TBE (44.5 mM Tris, 44.5 mM boric acid, 5 mM EDTA), which had been prerun at 4 C for 30 min at 100 V. The gel was run at 100 V in 0.5 x TBE, first for 20 min at 4 C and then for an additional 160 min at room temperature. Cooling the samples for a minimum of 10 min followed by electrophoresis into the gel at 4 C maximized the amount of specific STAT5-DNA gel-shift complex and minimized the formation of more rapidly migrating nonspecific complexes. After electrophoresis of the samples into the gel for 20 min at 4 C, the gel apparatus was moved to room temperature to increase the speed of protein migration. Cold probe competitions were carried out using up to a 200-fold molar excess of unlabeled DNA probe added in a volume of 1 µl just before addition of the ³²P-labeled DNA probe. For supershift analysis, antibody was added 10 min after the labeled DNA probe. Antibodies used were anti-STAT5b (Santa Cruz Biotechnology, Santa Cruz, CA; sc-835), anti-STAT3 (Santa Cruz, sc-482), anti-STAT1 (Transduction Labs, Lexington, KY; G16920), as described previously (13). Gels were exposed to phosphorimager plates for 16 h followed by quantitation using a Molecular Dynamics PhosphorImager and ImageQuant software (Sunnyvale, CA).

The STAT5/mammary gland factor response element of the rat ß-casein promoter (nucleotides -101 to -80), 5'-GGA-CTT-CTT-GGA-ATT-AAG-GGA-3' (sense strand) was used for EMSAs. The probe was end labeled with ³²P on one strand, annealed to the antisense strand, and then gel purified. STAT1 and STAT3 DNA-binding activity was assessed using a c-fos gene SIE probe as described elsewhere (13).

Immunoprecipitation, Western Blot, and Other Analyses
Immunoprecipitation with anti-phosphotyrosine antibody PY-20 and treatment with the phosphatases PTP1B and PP2A were as described (13). For Western blotting, total cell extracts were electrophoresed through Laemmli SDS-polyacrylamide gels (7.5% gels or 5–10% gradient gels) run at constant current and a starting voltage of 75 V, with cross-over to constant voltage at 170 V. In some cases (Fig. 3B), 6.5% gels were run overnight at a constant current of 6 mA to maximize resolution of STAT5b bands 1 and 1a. Gels were electrotransferred to nitrocellulose and probed with anti-STAT5 antibodies. Blocking and probing conditions were as described (12). Detection on x-ray film was by enhanced chemiluminescence using Amersham ECL reagents (Amersham, Arlington Heights, IL). Anti-STAT5a antibody was from Santa Cruz (catalog no. sc-1081). Anti-STAT5b antibody (Santa Cruz, catalog no. sc-835) was used in all experiments shown unless noted otherwise. This antibody was shown to be STAT5b-specific under our Western blotting conditions by analysis of extracts of Cos 1 cells transiently transfected with either mouse STAT5a cDNA or mouse STAT5b cDNA, kindly provided by Dr. Alice Mui (DNAX Corp., Palo Alto, CA) (data not shown) (53). The unphosphorylated, cDNA-expressed STAT5a migrated distinctly slower than STAT5b, i.e. at a position close to the GH-activated, diphosphorylated STAT5b band 2 (see Results), whereas cDNA-expressed STAT5b comigrated with the major STAT5 form present in untreated CWSV-1 cells, supporting our identification of the GH-activated STAT5 protein presented in CWSV-1 cells as STAT5b.

View larger version (K): [in this window] [in a new window]   Figure 7A. ence of band 2 in a similar experiment shown in Fig. 5B, lanes 10 and 11). C, The STAT5b-DNA-binding activity remaining 1 h after termination of GH pulse 1 was quantitated for each of the drug treatments shown and is expressed as a percent of the STAT5b activity determined 45 min after GH addition, as depicted in the diagram. Pervanadate, H7, and H8 were strongly effective in prolonging the STAT5 signal. Error bars correspond to SD values based on n = 2 to 6 independent determinations, except for calyculin A, HA1077, and KT5720, which are each based on a single experiment.

	ACKNOWLEDGMENTS

	FOOTNOTES

 
Address requests for reprints to: David J. Waxman, Department of Biology, Boston University, 5 Cummington Street, Boston, Massachusetts 02215.

Supported in part by NIH Grant DK-33765 (to D.J.W.).

¹ As shown in Fig. 3B, below, STAT5b band 1a (serine/threonine-phosphorylated form) migrates somewhat slower than STAT5b band 1 (tyrosine-phosphorylated form) on Western blots of SDS gels. In most cases, however, these two bands could not be distinguished electrophoretically and are therefore marked as ‘band 1/1a’ on each of the figures. In untreated cell extracts (e.g. Fig. 1B, lane 1) this band actually corresponds to the serine-threonine-phosphorylated band 1a, whereas in GH-stimulated extracts (e.g. Fig. 1B, lane 2), it corresponds to a mixture of band 1 and band 1a.

² The somewhat lower STAT5b DNA-binding activity and lower level of STAT5b band 2 observed after the second GH pulse as compared with the first GH pulse in this experiment and in the one shown in Fig. 7 may reflect the 2.5-h interpulse interval used in this experiment, which is shorter than the optimal 3-h interval identified in Fig. 6A.

Received for publication October 4, 1996. Revision received December 24, 1996. Accepted for publication December 30, 1996.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 

1. Isaksson OG, Eden S, Jansson JO 1985 Mode of action of pituitary growth hormone on target cells. Annu Rev Physiol 47:483–499[CrossRef][Medline]
2. Darnell Jr J, Kerr IM, Stark GR 1994 Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins. Science 264:1415–1421[Abstract/Free Full Text]
3. Ihle JN 1996 STATs: signal transducers and activators of transcription. Cell 84:331–334[CrossRef][Medline]
4. Levy DE 1995 Interferon induction of gene expression through the Jak-Stat pathway. Semin Virol 6:181–189
5. Ivashkiv LB 1995 Cytokines and STATs: how can signals achieve specificity? Immunity 3:1–4[CrossRef][Medline]
6. Finidori J, Kelly PA 1995 Cytokine receptor signalling through two novel families of transducer molecules: Janus kinases, and signal transducers and activators of transcription. J Endocrinol 147:11–23[Medline]
7. Argetsinger LS, Campbell GS, Yang X, Witthuhn BA, Silvennoinen O, Ihle JN, Carter-Su C 1993 Identification of JAK2 as a growth hormone receptor-associated tyrosine kinase. Cell 74:237–244[CrossRef][Medline]
8. Meyer DJ, Campbell GS, Cochran BH, Argetsinger LS, Larner AC, Finbloom DS, Carter-Su C, Schwartz J 1994 Growth hormone induces a DNA-binding factor related to the interferon-stimulated 91-kDa transcription factor. J Biol Chem 269:4701–4704[Abstract/Free Full Text]
9. Silva CM, Lu H, Weber MJ, Thorner MO 1994 Differential tyrosine phosphorylation of JAK1, JAK2, and STAT1 by growth hormone and interferon-gamma in IM-9 Cells. J Biol Chem 269:27532–27539[Abstract/Free Full Text]
10. Campbell GS, Meyer DJ, Raz R, Levy DE, Schwartz J, Carter-Su C 1995 Activation of acute phase response factor (APRF)/Stat 3 transcription factor by growth hormone. J Biol Chem 270:3974–3979[Abstract/Free Full Text]
11. Gronowski AM, Zhong Z, Wen Z, Thomas MJ, Darnell Jr JE, Rotwein P 1995 In vivo growth hormone treatment rapidly stimulates the tyrosine phosphorylation and activation of Stat3. Mol Endocrinol 9:171–7[Abstract]
12. Waxman DJ, Ram PA, Park SH, Choi HK 1995 Intermittent plasma growth hormone triggers tyrosine phosphorylation and nuclear translocation of a liver-expressed, Stat 5-related DNA binding protein. Proposed role as an intracellular regulator of male-specific liver gene transcription. J Biol Chem 270:13262–13270[Abstract/